Systematic Review of Deterministic and Rule-Based Models for QoS-Aware 5G Network Slice Classification

Abstract

The development of 5G technology has brought about network slicing as a key architectural advancement, allowing multiple virtual networks to function over a single shared physical infrastructure. Accurate classification of traffic into suitable slices enhanced Mobile Broadband (eMBB), Ultra-Reliable Low-Latency Communications (URLLC), and massive Machine-Type Communications (mMTC) is essential to ensure Quality of Service (QoS) and efficient resource utilization. While recent research has largely focused on machine learning and deep learning techniques, challenges such as lack of explainability, high computational cost, and limitations in real-time implementation have renewed attention toward deterministic, rule-based methods. This study conducts a systematic conceptual review of rule-based prediction models for 5G network slicing classification using the PRISMA framework. A comprehensive search of peer-reviewed studies published between 2022 and 2025 was performed across major academic databases. After undergoing identification, screening, eligibility evaluation, and final inclusion processes, a total of 30 relevant studies were analyzed. The results show that rule-based models offer advantages such as interpretability, low-latency decision-making, support for regulatory compliance, and strong suitability for deployment in edge computing environments. Based on these findings, a structured Rule-Based Prediction Model (RBPM) framework is proposed. The study concludes that rule-based approaches remain highly valuable for mission-critical 5G slicing applications and recommends their integration with adaptive techniques to enhance scalability and performance.